Major Service Outages — June 2026

A complete roundup of the biggest service disruptions in June 2026 — led by the June 22 Zayo fiber cut that took Reddit, Zoom, Discord, and Canva offline for hours at a time.

By Statusfield Team · Engineering··7 min read

June 2026 had one major headline outage and several smaller but significant incidents. Here's a month-end accounting of what broke, when, why, and what it means if your team depends on any of these services.


Incidents at a Glance

DateService(s)DurationRoot Cause
June 10–11Cloudflare R2, Images, Pages~18 hoursInfrastructure issue (partial)
June 10GitHub (API / Actions)~45 minSporadic auth failures
June 17Comcast (backbone)29 minRouting issue
June 19AT&T (backbone)45 minRouting issue
June 22Reddit, Zoom, Discord, Canva + more~4 hoursZayo fiber cut
June 23GitHub Copilot Completions~44 minInference degradation

The Big One: June 22 — Zayo Fiber Cut

The largest outage of the month — and one of the more instructive ones of 2026 — wasn't caused by a software bug or a misconfigured deployment. It was a fiber optic cable break.

What happened

At approximately 11:00 AM ET on June 22, a fiber cut in eastern North America severed routes on Zayo's backbone network. Zayo is a major internet transit provider — one of the invisible highways that carries traffic between data centers and end users. When its routing failed, any platform routing through those paths lost reachability.

The knock-on effects were immediate and wide:

  • Reddit — went dark globally; over 35,000 Downdetector reports at peak
  • Zoom — meetings failing across North America
  • Discord — voice and messaging degraded
  • Canva — design platform inaccessible for many users
  • Cloudflare — elevated latency and errors across Cloudflare-proxied sites
  • Microsoft Teams, Fortnite, Robinhood — also affected

Cloudflare identified the root cause at 14:44 UTC and declared the incident largely resolved at 17:30 UTC — roughly four hours after the initial break.

The status-page gap

This is the same pattern every major outage follows: Downdetector and user reports spike before any official acknowledgment. On June 22, Reddit, Discord, and Zoom users were reporting problems in the tens of thousands before any of the affected services had updated their status pages.

The reason is structural: status-page updates require a human to notice, triage, write, and post. User reports require none of those steps.

If you found out about the June 22 outage because your users told you — that's the gap. Monitoring closes it.

Why a fiber cut affects software companies

This is the part most engineering teams don't internalize until they've lived through it: your service's uptime is bounded by your transit provider's uptime, and you have no visibility into that layer.

Every major cloud provider (AWS, GCP, Azure), CDN (Cloudflare, Fastly, Akamai), and SaaS platform has at least some traffic routing through backbone providers like Zayo, Cogent, Lumen, and NTT. A break at that layer can affect dozens of services simultaneously — not because any of them failed, but because the roads between their servers and your users failed.

The June 22 outage is a good example of why monitoring matters across your stack, not just for the services you use directly.


Cloudflare: R2, Images, and Pages (June 10–11)

Cloudflare's object storage (R2), image resizing, Images uploads, and Pages deployments experienced degraded availability across June 10 and into June 11. This wasn't a full outage — Cloudflare's core CDN and proxying remained operational — but teams using R2 as a media storage layer or deploying via Pages were blocked.

Who felt it: Development teams with CI/CD pipelines through Cloudflare Pages, and applications storing user-uploaded images in R2.

Duration: Partial degradation over approximately 18 hours before full resolution.

Check current Cloudflare status →


GitHub: API Auth Failures (June 10)

On June 10, GitHub mitigated sporadic authentication failures affecting approximately 15% of API traffic. The incident ran for roughly 45 minutes and primarily hit API consumers — CI/CD pipelines using GitHub tokens, integrations that poll the GitHub API, and GitHub Actions workflows that authenticate mid-run.

Who felt it: Any automated workflow or integration that uses GitHub's API. Actions pipelines were the most common failure mode — a run would start, hit an auth failure at a token exchange, and fail the whole build.

On June 23, a separate GitHub Copilot incident hit between 22:45 and 23:29 UTC: Completions and Next Edit Suggestions failed for approximately 25% of requests. That incident ran for 44 minutes.

Check current GitHub status →


What June 2026 Taught Us

Three patterns worth noting across this month's incidents:

1. Infrastructure dependencies are the hardest layer to monitor yourself. The Zayo fiber cut took down platforms from Discord to Canva — companies with excellent engineering teams and solid internal infrastructure. Their uptime was limited by a provider they don't operate and can't observe directly. The only way to know any of those services are down is to watch their observable status, not their internal health.

2. Status pages lag reality. Every incident this month had a gap between first user reports and first status-page acknowledgment. That gap is unavoidable — status pages require humans to update them. For the services your team depends on, that gap is your blind spot.

3. Fiber cuts, routing failures, and backbone incidents are not rare. June had at least three (Zayo June 22, Comcast June 17, AT&T June 19). These are outside any single company's control but entirely within your detection range if you're monitoring the right layer.


Track These Services on Statusfield

If any of the services above are in your stack, you can monitor their status directly:

Statusfield checks each service every 5 minutes and alerts you via email, Slack, Discord, or webhook the moment status changes — in either direction. 3 monitors, 10 alerts/month, no credit card required. Start monitoring →


FAQ

What was the biggest outage in June 2026? The June 22 Zayo fiber cut. It took Reddit, Zoom, Discord, Canva, Cloudflare, and several other major platforms offline simultaneously for approximately four hours. The root cause was a physical fiber optic cable break in eastern North America affecting Zayo's backbone routing.

How long was the June 22 outage? Cloudflare identified the root cause at 14:44 UTC and declared the incident largely resolved at 17:30 UTC — roughly four hours from when the disruption began (~11:00 AM ET / 15:00 UTC).

Was Cloudflare down in June 2026? Twice. On June 10–11, Cloudflare R2, Images, and Pages deployments experienced degraded availability over approximately 18 hours. On June 22, Cloudflare's network was also impacted by the Zayo fiber cut, with elevated latency and errors across Cloudflare-proxied services.

Was GitHub down in June 2026? Yes, twice. On June 10, sporadic API authentication failures affected approximately 15% of API traffic for ~45 minutes. On June 23, GitHub Copilot Completions experienced ~25% request failures between 22:45 and 23:29 UTC.

How do I know when Reddit or Discord are down? Statusfield monitors their official status channels and alerts you the moment any component changes status. Set up monitoring →


Published: July 1, 2026. Covers incidents June 1–30, 2026. Check live service status →

Know the moment a tool you depend on goes down

Statusfield watches 7,000+ services your business depends on and alerts you the moment they break.

Free plan · No credit card